tojacobs,
You are correct: The CRUNCH (version 1 and/or version 2) is a tool. Here my thoughts:
There are in a way two (almost) separate approaches of using the tool.
Approach 1: If you have only one source of data (or maybe two) your requirement is to reduce the input data to a playable amount by using the built-in filters the CRUNCH provides.
The reudcton part of the CRUNCH is what you need. The CRUNCH 2 has currently 120 filter selections, from basic High/Low and Even/Odd, Sums and Roots filters to more sophisticated filters such as Vtracs, Sedertree's Spike/Trough, more specific Hih/Medium/Low filters and more.
You can see examples of the one-source use in many of Lucky's recent posts, where he uses one source and reduces the numbers through sums and key numbers particularly. In this case high priorities are the most likely picks. As another example Latern is using more specific filters in his posts, but basically uses this approach as well.
Approach 2: If you have multiple sources of data, this changes you focus from reducing to combining. Let's say you are using different predictors or different programs to get your data. You enter these lists into the CRUNCH. Even without much filtering you will get usable results.
How? You will use the priorities in a different way than in the Approach 1. Depending on how many soruces you are using you will look for teh best numebrs to play in certain priority groups. Usually you are looking for the area that paymenow calls "midpoints", the priorities that are not the highest and not the lowest.
Why is that? You are trying to find the numbers that from all your sources end up in the middle, so to speak the "average" numbers. For example if you have equal 5 sources of data you are looking for numbers that occur between 2 and 3 times. If you have 10 sources you are looking for numbers that occur about 5 times (betwen 4 and 6 times maybe). You will find this approach in some of my posts, in paymenow's posts.
These are basically my thoughts concerning the usage of CRUNCH for predictions, and the expected usage of this tool, besides creating all kind of charts and fequency statistics.
GOOD LUCK!